Technical Field
The present disclosure relates to a differential amplifier circuit for a capacitive acoustic transducer.
Description of the Related Art
As is known, an acoustic transducer of a capacitive type generally comprises a sensing structure, designed to transduce acoustic pressure waves into an electrical quantity (in particular, a capacitive variation), and an electronic reading interface, designed to carry out appropriate processing operations (amongst which amplification operations) on said electrical quantity for supplying an electrical output signal (for example, a voltage).
The sensing structure in general comprises a mobile electrode, in the form of a diaphragm or membrane, arranged facing a fixed electrode, at a short distance (the so-called “air gap”), to form the plates of a sensing capacitor with a capacitance varying as a function of the acoustic pressure waves to be detected. The mobile electrode is generally free to move, or undergo deformation, in response to the pressure exerted by the incident acoustic waves, in this way causing the capacitance variation of the sensing capacitor.
For example, MEMS (Micro-Electro-Mechanical System) capacitive acoustic transducers are known, in which the sensing structure is of a micromechanical type built using integrated micromachining techniques typical of the semiconductor industry.
By way of example, FIG. 1 shows a micromechanical structure 1 of a MEMS acoustic transducer, of a known type, which comprises a structural layer or substrate 2 of semiconductor material, for example silicon, in which a cavity 3 is made, for example via chemical etching from the back. A membrane, or diaphragm, 4 is coupled to the structural layer 2 and closes the cavity 3 at the top. The membrane 4 is flexible and, in use, undergoes deformation as a function of the pressure of incident acoustic waves.
A rigid plate 5 (generally known as “back-plate”) is arranged facing the membrane 4, in this case above it, via interposition of spacers 6 (for example of insulating material, such as silicon oxide). The rigid plate 5 constitutes the fixed electrode of a sensing capacitor with variable capacitance, the mobile electrode of which is constituted by the membrane 4, and has a plurality of holes 7, designed to enable free circulation of air towards the membrane 4 (rendering the rigid plate 5 in effect acoustically transparent).
The micromechanical structure 1 further comprises (in a way not illustrated) electrical membrane and rigid-plate contacts, used for biasing the membrane 4 and the rigid plate 5 and acquiring a sensing signal indicative of the capacitive variation due to deformation of the membrane 4 caused by incident acoustic pressure waves. Typically, these electrical contacts are arranged in a surface portion of the die in which the micromechanical structure is provided.
In general, the sensing structure of capacitive acoustic transducers is charge-biased, usually via a fixed charge. In particular, a DC biasing voltage is applied, usually from a charge-pump stage (the higher this voltage, the greater the sensitivity of the microphone), and a high-impedance element is inserted (with impedance of the order of teraohms, for example between 100 GΩ and 10 TΩ) between the charge-pump stage and the sensing structure.
This high-impedance element may, for example, be provided by a pair of diodes arranged in back-to-back configuration, i.e., connected together in parallel, with the cathode terminal of one of the two diodes connected to the anode terminal of the other, and vice versa, or by a series of pairs of diodes in the back-to-back configuration. The presence of this high impedance “insulates” the DC charge stored in the sensing structure from the charge-pump stage, for frequencies higher than a few hertz.
Since the amount of charge is fixed, an acoustic signal (acoustic pressure), which impinges upon the mobile electrode of the sensing structure, modulates the gap with respect to the rigid electrode, producing a capacitive variation and consequently a voltage variation.
This voltage is processed, in the electronic interface, by an electronic amplifier circuit, which is required to have a high input impedance (to prevent perturbation of the charge stored in the micromechanical structure), and then is converted into a low-impedance signal (designed to drive an external load).
FIG. 2a shows a possible embodiment of the amplifier circuit, designated by 10, in this case with a single output, of a so-called “single-ended” type.
The sensing structure of the capacitive acoustic transducer, designated as a whole by 11, is represented schematically by a sensing capacitor 12a with capacitance CMIC, which varies as a function of the acoustic signal detected, connected in series to a voltage generator 12b, which supplies a sensing voltage VSIG (having, in the example illustrated in FIG. 2, a generically sinusoidal waveform).
Given that, typically, the mobile electrode has a high parasitic capacitance towards the substrate (comparable with the capacitance of the sensing capacitor of the sensing structure), whereas the rigid electrode has a lower parasitic capacitance, the mobile electrode is generally electrically connected to a first low-impedance input terminal N1, for example to a reference ground voltage of the circuit, whereas the rigid electrode is electrically connected to a second input terminal N2, on which the sensing voltage VSIG, indicative of the capacitive variations of the sensing capacitor 12a, is acquired.
The second input terminal N2 is further electrically connected to a biasing stage, for example a charge-pump biasing stage (here not illustrated), by interposition of a first high-impedance insulating element 13, constituted by a pair of diodes arranged in back-to-back configuration, for receiving a biasing voltage VCP.
The amplifier circuit 10 further comprises a decoupling capacitor 14, and an amplifier 15, in buffer or voltage-follower single-ended configuration (i.e., with its inverting input connected to the single output). For example, the amplifier 15 is a class-A, or else a class-AB, operational amplifier.
The decoupling capacitor 14 (which operates to decouple the DC component and couple the detected signal) is connected between the second input terminal N2 and the non-inverting input of the amplifier 15, which further receives an operating voltage VCM from an appropriate reference-generator stage (here not illustrated), via interposition of a second high-impedance insulating element 16, constituted by a respective pair of diodes arranged in back-to-back configuration.
The operating voltage VCM is a DC biasing voltage, appropriately selected for setting the operating point of the amplifier 15. This operating voltage VCM is chosen, for example, in a range comprised between a supply voltage of the amplifier 15 (not shown) and the reference ground voltage.
During operation of the capacitive acoustic transducer, the (AC) sensing voltage VSIG is thus superimposed on this DC operating voltage VCM.
The amplifier 15 supplies on the single output OUT an output voltage VOUT, as a function of the sensing voltage VSIG detected by the sensing structure 11 of the capacitive acoustic transducer. The output voltage VOUT has, in the example, a sinusoidal waveform that corresponds in amplitude to the sensing voltage VSIG (as represented schematically in FIG. 2a).
FIG. 2b shows a further embodiment of a known-type amplifier circuit 10, which also in this case has a single-ended output.
The amplifier circuit 10 here comprises a MOS transistor 17 (in the example of a PMOS type) in source-follower configuration, which has its gate terminal connected to the second input terminal N2 via the decoupling capacitor 14, its source terminal that supplies the output voltage VOUT on the single output OUT, and its drain terminal connected to the reference ground voltage.
The source terminal of the MOS transistor 17 further receives a biasing current IB from a current generator 18 connected to a line set at a supply voltage Vcc. The second insulating element 16 in this case couples the gate terminal of the MOS transistor 17 to the ground reference.
In general, the single-ended circuit configuration presents some drawbacks, amongst which a poor rejection in regard to any common-mode component of disturbance, for example, deriving from power-supply noise or cross-talk from nearby devices with time-varying signals.
To overcome these drawbacks, it has been proposed to replace the single-ended solution with a configuration, defined as “pseudo-balanced” or “pseudo-differential”, which is illustrated in FIGS. 3a and 3b. 
This solution envisages that the amplifier circuit 10 comprises a dummy capacitor 19, constituted by a capacitor of a classic type, for example metal-oxide-metal (MOM) or metal-insulator-metal (MIM), having a capacitance CDUM, with a nominal value substantially equal to the capacitance at rest (i.e., in the absence of external stresses) CMIC of the sensing capacitor 12a. 
The amplifier circuit 10 has in this case the exact duplication of the circuit elements previously described with reference to FIGS. 2a and 2b (the duplicated elements are distinguished with a prime sign in FIGS. 3a and 3b and are not described again), for generation on a further output OUT′ of a DC output voltage Vout _DUM designed to balance the output voltage Vout, thus enabling elimination of the common-mode disturbance. Basically, two altogether equivalent circuit paths are created.
Also this solution is, however, not free from drawbacks.
In particular, given that the contribution of the sensing signal is present only on one of the two circuit paths, i.e., the one that goes from the sensing capacitor 12a to the output OUT (thereby, the “pseudo” differential nature of the amplifier circuit 10), on the same output OUT a greater voltage swing is required, in particular with a value twice that of a fully differential solution (where half of the swing would be present on the output OUT and the other half of the swing, in phase opposition, on the further output OUT′).
A higher value of the supply voltage Vcc is thus required, with consequent increase in power consumption.
To overcome the above problem related to the swing on the output of the amplifier, a further solution that has been proposed, illustrated in FIG. 4, envisages use of a differential amplifier 25 with four inputs and two outputs, the so-called DDA (Differential Difference Amplifier), having a differential and unity-gain architecture (the voltage difference between the differential output terminals, which are here designated by Out1 and Out2, is equivalent to the voltage difference between the differential input terminals, which are here designated by 25a and 25c).
The structure of the differential amplifier 25 is described in detail for example in:
“A versatile building block: the CMOS differential difference amplifier”, E. Sackinger, W. Guggenbuhl, IEEE Journal of Solid-State Circuits, Vol. 22, April 1987; or
“A CMOS Fully Balanced Differential Difference Amplifier and Its Applications”, H. Alzaher, M. Ismail, IEEE TCAS-II: Analog and Digital Signal Processing, Vol. 48, No. 6, June 2001.
In particular, the second input terminal N2 is in this case connected, via interposition of the decoupling capacitor 14, to a first non-inverting input 25a of the differential amplifier 25, a first inverting input 25b of which is directly feedback-connected to a first differential output terminal Out1.
Likewise, the dummy capacitor 19 has a respective first terminal, designated by N1′, connected to the ground reference terminal, and a second terminal N2′ connected, via interposition of a respective decoupling capacitor 14′, to a second inverting input 25c of the differential amplifier 25, a second non-inverting input 25d of which is further directly feedback-connected to a second differential output terminal Out2 (the output voltage Vout being present between the first and second differential output terminals Out1, Out2).
The respective second input terminal N2′ of the dummy capacitor 19 further receives the biasing voltage VCP through a respective first insulating element 13′, constituted by a pair of diodes arranged in back-to-back configuration and receiving the biasing voltage VCP. Likewise, the second inverting input 25c receives the operating voltage VCM, via a respective second high-impedance insulating element 16′, which in the example is also constituted by a pair of diodes arranged in back-to-back configuration (the operating voltage VCM is thus a common-mode voltage for the first non-inverting input 25a and for the second inverting input 25c of the differential amplifier 25).
The dummy capacitor 19, in this case, enables creation of a substantially balanced path for the buffer inputs (i.e., the non-inverting input 25a and the inverting input 25c) of the differential amplifier 25, for a better common-mode noise rejection.
Even though it enables improvement of the capacity of disturbance rejection, also the differential configuration described with reference to FIG. 4 has some drawbacks.
In particular, the above solution involves two differential input stages, with a consequent increase in noise and current consumption. It has a wide common-mode input interval on account of the different signals present on the four inputs of the differential amplifier 25. The transistors of each input stage are driven by a differential signal having a large amplitude, i.e., the virtual-ground principle does not apply (as is known to a person skilled in the sector), with a consequent high distortion for signals of large amplitude. Finally, the differential signal is effectively present only on the differential output terminals Out1, Out2, whereas the input terminals are not fully differential.
In general, the need is thus felt to provide an amplifier circuit for a capacitive acoustic transducer that will enable the disadvantages and limitations associated to known solutions to be overcome, at least in part.